I'm really happy with how the number of bug reports coming in is not a massive deluge of different bugs, but mostly just endless repetition of the same handful. For some of them (like the notifications collapse triggered crash on exit) we're getting several duplicates per day, so we know it's not just because people aren't using or reporting!
Even better than having bugs reported is having bugs fixed, of course. Which is exactly what's been happening in droves. From multi-screen fixes to panel hiding fixes to, yes, fixing that on-exit crash, we're smacking the bugs down as fast as we can so that next month's 4.6.1 release will be beautifully solid.
I'm hoping that the upcoming springtime releases of the various distributions will be able to benefit from this bug fixing spree as well so that users installing the defaults will have a great experience too.
Not that 4.6 is the only thing we're on to these days, of course. I've put together an activities runner (and speaking of activities, here's another nice blog entry on them, this time from Hans Chen) so that in 4.7 you can type thing like "activity" in KRunner to get a list of activities or follow that up with the name of (or start of the name of!) an activity to switch to it. Marco's been working on some Plasma application dashboard love and more QML goodies. We have "Apply" buttons in the Plasmoid config dialogs (shock! shock!). We've also begun in earnest on our 2011 roadmap (well, it actually stretches into the beginning of 2012, but.. details :) and started planning for our next dev sprint, Tokamak V, in the spring.
One thing we already know is that we feel we're on the "right" track with activities now and will spend a significant investment of our resources in 4.7 to fleshing them out further.
Meanwhile, I'm counting down the days until the movers come to grab my stuff to put on a dock so it can go in a boat which can go to the UK which can then drop it on another dock before it somehow, somewhen makes its way to my doorstep in Zurich. I leave here on the last day of the month, and there just doesn't seem to be enough days left. A little nerve wracking ... but very exciting! I can't wait to be in the new place, meeting new people and, of course hacking on good ol' KDE from a new location. One where I don't have to be up at 1:00 am on a regular basis just to attend team meetings. :)
Subscribe to:
Post Comments (Atom)
30 comments:
Off topic: There's a question I've been wondering about regarding KDE for a while, if you wouldn't mind answering it. That is, why did you and the other KDE hackers decide to release KDE 4.0 and 4.1, instead of, say, following the Debian philosophy of "we'll release it when it's ready"? You could have just left us with KDE 3.5.x until you were done with KDE 4.2 (which AFAIK was widely regarded as the first "usable" version of KDE).
KDE in general, and you in particular, got bashed pretty hard over KDE 4.0, too.
So why didn't you hold back on releasing it?
- Andrew H.
@Andrew: in a phrase, "release early, release often."
if we didn't release then, we would have sat for another year or more not getting enough testing of the sort the matters while people continued to make all sorts of destabilizing changes.
freezing for longer than that with not user available release would mean stagnating. fewer users (increasingly so over time), fewer developers (who get bored or stop believing there will be a release).
instead, we made a release to mark binary compatibility and the start of getting serious about whipping it all into shape. this is how open source used to be done. you know, silly projects like Linux and Apache. ;) but this knowledge has increasingly been lost as the f/oss community grew and became populated by more "average" users as well as more technically shallow (and at times outright shameless) self-promoters.
but had we not release 4.0, we wouldn't have 4.6 today. in fact it is highly doubtful we'd be anywhere near as far along. KDE may have been dead in the water.
look at GNOME3. they keep delaying. has it helped any? nope.
was the decision easy? no. was living with it easy? no. but we knew what we were doing and stuck to it.
"KDE in general, and you in particular, got bashed pretty hard over KDE 4.0, too."
which says more about those people who did the bashing than KDE or myself, i think. :)
You did a good job at staying positive and on-topic while some very negative things were being said. It was sometimes painful to watch.
Thanks for the reply. As an end-user, I must admit I used the occasion of KDE 4.0 to get to know GNOME better, but KDE is excellent.
(As far as GNOME 3 goes -- I'm typing this from a GNOME 3 preview CD, and I just don't "get" the utility of a UI composed of a large window filled with a lot of icons that covers most of the desktop. One is apparently supposed to search for the icon of the application one is looking for from amongst them, or use the search bar. What was wrong with nicely organized cascading menus? Sigh. Change isn't always for the better. Also, it sports a KDE4-like "system settings" applet.)
@Andrew:
Many applications from 4.0 were already held back because the Plasma devs worked hard to get in a dot-0 release shape (which it was and whoever claims that a dot-0 release has to be flawless lives in a 'pink ponies and fairies' world).
Holding back the whole release just because a single component was not perfect, would have been very unfair to all application developers.
openSUSE 10.3 installed quite a bunch of SC4 application with KDE3.5, even tough 4.0 was officially still in pre-release state.
One of the worst examples of your 'release never, delay often' approach is Enlightenment. Their latest version is in development for over 10 years. I'm not kidding. During the time of KDE2 (the time when E17 development started) Enlightenment was considered a serious competitor to KDE.
Even worse is Hurd whose devs delay its release by one year since the 1980s...
"One thing we already know is that we feel we're on the "right" track with activities now and will spend a significant investment of our resources in 4.7 to fleshing them out further."
This sounds really great!
I myself have finally started using activities with 4.6, and they really do increase productivity.
I believe that once they're polished just a little further (i.e. made even easier to create, manage and switch between, and maybe come with some nice defaults and hints (discoverability!) for new users), this could become a "killer feature" that might actually convince people to switch to KDE from, let's say, Windows 7.
It's great to see that KDE is leading innovation here, rather than just trailing behind the commercial competitors!
PS: I also REALLY enjoy the integrated Nepomuk search feature in Dolphin in 4.6... (And of course also the exceptional overall stability and performance of this release... Great job!)
Oo! Oo! Can I make a suggestion for activities?
Let us manage them in the K-menu. Pleeeeease?
Is there a way to get all KDE bugs as a RSS feed what would be updated in realtime?
To get just the subjects?
I think that would help even many bug hunters and avarage users to easier way to notice what is being reported so they would not report same bugs (the search does not always find anything similar).
From Andrew:
(As far as GNOME 3 goes -- I'm typing this from a GNOME 3 preview CD, and I just don't "get" the utility of a UI composed of a large window filled with a lot of icons that covers most of the desktop. One is apparently supposed to search for the icon of the application one is looking for from amongst them, or use the search bar. What was wrong with nicely organized cascading menus? Sigh. Change isn't always for the better. Also, it sports a KDE4-like "system settings" applet.)
Try out the plasma-netbook shell on a tiny netbook screen, and you'll come to love having "the desktop" being "the launcher". Haven't tried GNOME Shell or Unity or anything like that, but the way plasma-netbook works is a thing of beauty.
Not too useful with resolutions above 1024x768, though.
I have it running on 7" and 10" netbooks, and my laptop "media centre" connected to the TV via S-Video (800x600).
Screens that small, you really don't want to be scrolling around endless menus. :)
Given this information:
http://www.nytimes.com/2011/02/04/technology/04nokia.html
do you think, Qt (and hence KDE) is up for rough days ahead? May be pulling off a "LibreOffice" is on the cards?
I am glad things are looking up. I think that Gnome is doing the right thing by not releasing Gnome 3 until it is ready. I for one will not use it like I will not use KDE any longer just because it no longer does what I need it to do. The older I get the less I need all the bells and whistles and glad that I have alternatives that will give me this. I hope that things keep improving and hopefully you will fix more bugs before releasing a version that has bugs in it. I realize that more bugs are getting fixed each time but I for one would rather wait a while before releases and have a solid product instead one just released to release.
Good. Great.
Now, figure out how to make Akonadi and This whole 'semantic desktop' thing an uninstallable option, and I'll take another look.
I have to say I´d rather see some basic features getting more attention instead of trying to come up with something as complicated as activities (I say so because many people still don´t get the concept, and there´s a piece of pure feedback).
I still think KDE is inconsistent at times (sports a configuration center but not all things configuration are on it, i.e. KDE menu, custom shortcuts, etc.), over-complicated where it shouldn´t be (does anybody truly use all features available in color theming?) and unreliable where it counts (could we please get some solid icon theming working already? I think it´s about time, specially for a project that supposedly considers Look&Feel so important.)
On a different note, looking at the revolution the cloud is bringing and how new devices like smartphones and tablets are changing the way users understand the GUI, is it really a good idea to invest time and resources into creating an even thicker and more complicated desktop, then expect users to go through a learning curve?
GNOME3 and Unity may not be my cup of tea, but I think those guys have realised that the rules of the game are changing... Whether their interfaces will actually be successful is a completely different matter, though.
I found 4.0 a bit stressful, but that was my problem.
For the most part I've been pleased and I've always been grateful to both openSUSE and KDE.
I do feel those of us further down the food chain, using and bug reporting have been let down by the openSUSE/nVidia/KDE>4.5.5 problem.
It only hit my machine post 4.5.95, but the only information we've been given is that it's upstream. That doesn't feel logical.
For better or for worse, 'we' are part of your promotions team, and we could have been given a better explanation, not least because rolling back is such a pain but also because it would help us understanre remains bright.
I think I get the concept of activities but I don't get the implementation. I really hate the panelesque popup, personally I wish I could manage them with widgets... or something that I could put always visible in my panel area (or if I had to resize it) like I can with virtual desktops.
I also seem to recall you telling me that only 1 activity can be active at a time, yet I have seen the ghosted other activity widgets on my desktop, which makes no sense to me. And I can't assign applications to "stopped" activities. so I feel something in that nature is amiss.
oh yes.. nepomuk... bangarang + dolphin + nepomuk now finally useful... unfortunately bangarang is a bit slow and need some polish, but it'll come.
Zurich. Nice. Switzerland is my favourite holiday location. The views, the places you can go to, just beautiful. I'd find it harder to hack there because of it :)
@Aaron J. Seigo
Regarding early releases of KDE...
"The perfect is the enemy of the good" - Voltaire.
Not enough people beta test as it is. I agree, getting the updates out early is the way to go, otherwise as you said, stuff would stay in beta forever. That's more of a pain for most people, as it usually means the only way to try the software is by compiling it or jumping through some other hoops.
FINALLY an apply button for plasmoids. It seemed a bit "gnome-like" to just have close there. I wasn't sure my changes were being saved (at first) since KDE has apply everywhere else.
@Chema Martin: Cloud computing sucks. Tiny little mobile devices with screens you can barely see suck. I want everything on my hard drives on my LAN behind my firewall... and on screens big enough that I can see what I'm doing. Let Gnome with the Gnome Shell or Ubuntu with Unity try designing an interface essentially for smaller devices and then try ramming it down the throats of desktop users... I think they're going to fail. KDE is going in the right direction.
I would agree with the comment about activities. I think that conceptually there is not a very good separation between activities and virtual desktops so they remain somewhat confusing and ambiguous (in terms of what they are trying to accomplish.) I think that some thought needs to be put into streamlining this, making it more discoverable, easier to use, etc.
I think that what KDE needs now is some spit and polish. I agree with what Chema said. Consistency needs to be addressed, things need to be streamlined, etc. (E.G. The control center is a mess and it has been for a very long time.) I think a UI/Workflow designer/analyst (who is not a coder) needs to do a thorough analysis to ensure that workflow concepts and paradigms are consistent across the entire desktop. (Performance also needs to be addressed, not sure if that is better in 4.6 but KDE4 has generally been very sluggish compared to Gnome, it's what's kept me away.)
I am moving to Fluxbox or the now very very very stable enlightenment. I dropped KDE since version 4.5 and switched to gnome. I got tired of waiting for KDE to behave the way a Desktop environment should just work out of the box. Great job with the release often paradigm but like you lost Linus Torvalds you lost me too. As a final note, I hate to say this but after so long of using KDE 4.x I still don't get the concept and difference between activities and virtual desktops.
the one goes, others will come. I'm just preparing a patch to get KDE 4.6+ into the BLFS book. In the meanwhile (since 4.0) KDE became more and more stable and time is right to honor this. Now i think its the time were KDEs status changes from "incredible unstable/unusable" to "has some more or less heavy bugs to fix". Good job, keep going!
Beside this, there are some things to note. The monthly release cycle is what IMHO should be in focus. It really does not good to the reputation of KDE, not after the 4.0 desaster, if there is a release which contains really heavy bugs. Delay the release if more time is needed to fix. Noone will blame anyone when a release is delayed for bugfixing, I'm sure, aren't you?
Next is the kdesupport. What the hell was this idea? There are many tarballs containing all the KDE stuff, perfectly presented with checksums and such. And than? You'll need this package and that, this one relies on that and so on. Many of the prerequisites are required, at least "strongly recommended" (e.g. polkit-qt) but there isn't a package to download. Yes, this stuff is in kdesupport, but there is no tarball containing the sources on which a KDE release regards to. Some of the packages have standalone project pages somewhere else (why the hell are they replicated), other not. If KDE relies on particular versions, than release tarballs for that. If it does not, release tarballs as every other project does too. It makes live easier and you can talk about particular versions in case of bugs. Anyhow, I dont get the idea of that kdesupport stuff, maybe thats only my problem.
Finally I like to say hold off from adding new features and technologies to KDE but get it as rock solid as 3.5.10 was. The dependencies are enough, KDE is by far the heaviest pack you can install. Optimize it and get it stable and faster.
--
Thomas
@Supergeek,
Good luck with your new DE. When you miss KDE, it will still be there for you.
You don't have to be a super geek to understand Activities, but since you are, I am sure you understand the concept of Object Oriented concepts.
Now for the concept, just imagine a virtual object identified by a name. It has Properties and Methods/functions. Values of its properties could be files, folders, icons for applications, a Virtual Desktop/Page/Screen, etc... and for its methods, functions that could be triggered by an event. So, how hard is that hard to understand for a super geek?
"I still don't get the concept and difference between activities and virtual desktops."
There is resemblance between the two, but Activities are conceptual and more inclusive. they include VDs as part of their grouping. VDs are more of a visual grouping and limited. VDs in an actitity are associated with its sub-objects, i.e. an applications has a specific VD/Page associated with it, or a file/folder/etc when triggered and needs to dispaly something, it uses the specific VD associated with it, if it doesn't have one associated, it uses the VD that is associated with its parent, the activity, or may be a default one.
So who needs Activities? Well, if you have worked in business environment, you might have noticed that projects are pefect example to create activities for. Activities will simplify project information access, better maintain the data, and control record keeping.
Are activities for everyone? Not at all and they are one option to use or not use. So why are you bothered by them if you don't have to use them?
I have one off topic question which is buggin' me for some time.
When I heard about activities for the first time my first reaction was yes, exactly what I want. Different applications, different plasmoids, different wall paper, heck even different folder views. But I am still missing one little thing to make each activity really unique:
different taskbar/system tray and as an extra bonus even different kick-off application menu entries.
Version 4.6 was really excellent step forward when it comes to assigning specific application to specific activity.
Brilliant work guys and thank you .... but still waiting for things mentioned above. Any plans to make "activities experience" really unique?
"release early, release often"
There are a LOT of current Linux users who never experienced the early days of Linux or KDE and don't understand how a free project works. FOSS projects are a collaborative affair between the developers and the users. The user's part is to help test the apps and post bug reports. They can't do that if they have nothing to test! In exchange for helping they get the free use of superb software. And of the 30,000 or so applications in my Kubuntu's repository, KDE4 is far and away the best. The 4.5.3 release has given me the fastest, most powerful and most stable DE I have ever used since I began using a PC in the summer of 1978. I can't wait to upgrade to 4.6.
As far as Nioka switching from Symbian to Windows 7, why would they want to switch from a 31% market share to 3% market share?
As far as
@Gene: That's you, but most people will walk the cloud path, and it's no surprise.
At the rate it's going, the browser will be one of the most powerful applications not long from now. WebGL, HTML5, CSS3... Doors are opening for things that not long ago were bounded to thick clients. Fortunately, that is not the case anymore, allowing users to keep their data fully integrated across applications and devices, safe and resilient. Collaboration takes a whole new meaning, achieving things that are unthinkable in a traditional desktop environment. Also very importantly, a lot of these projects are open source, so it's no longer about lack of freedom.
If you have not already, I recommend taking a quick look at the Chrome web store. The amount of stuff that can be done online is insane, and once you get into it, lots of applications in the desktop just look arcane and redundant. It is true that I can simply ignore them, but it bugs me that time is put into those apps that will be getting less and less attention over time, when the stuff that really matters is still undone.
On a different note, privacy may be a big deal for you, as should be for everybody, but that shouldn't stop people from using cloud computing. One can use his/her brains to decide which data is sensitive, but in any case, how many people have not already uploaded data that could be considered sensitive into Gmail, Facebook, YouTube or Twitter?...
Trying to stop the cloud revolution is wasted time and effort. Instead, I think efforts should concentrate on getting the right laws in place so that privacy is respected.
To finish my post, just a thought: In the next 2-3 years, how many people do you know that would put up with the learning curve that KDE (or any traditional desktop) requires, it's lack of stability when new releases go live, its heaviness and slowness, when they have something like ChromeOS that boots in less than 7 seconds, requires no application knowledge (other than the browser) and gets them through to their information, media and games straight away?
I think a lot of people will use the cloud but for me I still want a desktop. I want to have my data local. I like the idea of having certain things in the cloud but some things I want to keep on my PC. As I get older my needs get more basic. This is why I am now using openbox as I do not need what KDE does any more. I hate how the applications are all being built with a lot of over head now. At one time they talked about being able to run the applications on any system but now you need a lot more just to run them. I do not find any thing all that great that I can not find else where with less dependencies.
@Supergeek
"I dropped KDE since version 4.5 and switched to gnome. I got tired of waiting for KDE to behave the way a Desktop environment should just work out of the box."
What are you talking about?
KDE wasn't really stable in 4.0 times, but 4.5, and especially 4.6, is stable and does provide a perfectly usable (and even fun-to-use) desktop environment out of the box.
If you still ran into problems, did you report them to the bug tracker?
Did you look in the forums for discussions on known issues?
Did you make sure you're not running an outdated buggy graphics driver (which seems to be one of the prime causes of problems users have with KDE these days)?
"I still don't get the concept and difference between activities and virtual desktops."
There have been several very good blog posts explaining the concept of "activities". If you're interested in the details, read them!
Here's how I would describe it:
Basically, KDE's 'activities' feature provides an easy way to instantly hide everything belonging to the real-life task you were pursuing up until now on your computer (e.g. report writing), and instead bring up everything belonging to another task (e.g. photo editing).
"Everything" in this context means Plasma configuration, plasma widgets, and application windows - and in the future, possibly other useful states (e.g. related to the global instant messaging system) as well. (Right, aseigo?).
And not only can you instantly switch between multiple running 'activities' (i.e. show/hide the associated widgets and windows and so on), but you can at any time stop an 'activity' (i.e. quit the associated windows/applications), and re-start an 'activity' (i.e. start and restore the windows/applications to the state they were left in when the 'activity' was stopped (as good as possible)).
To top it off, all of this is persistent across KDE sessions.
Virtual desktops, on the other hand, provide nothing more than some additional screen space.
Sure, they can be used to manually "simulate" some of the functionality provided by Activities (and commonly have so in the past, due to lack of alternatives), but only in a very limited fashion, and requiring a lot of manual "micro-management" of windows and screens.
@aseigo:
Maybe Activities would be better received in the KDE user community, if the default setup didn't include any virtual desktops and the default panel wouldn't include a pager, but instead some form of compact activity switching widget/button?
(Or maybe add an additional, auto-hiding panel on one of the screen edges in the default panel configuration, containing nothing but the (already available) activity switcher plasmoid?)
This could increase discoverability and avoid confusion for new users.
I was wondering if you guys are making use of Clang/LLVM and just spending some time fixing nasty looking compiler warnings and proactively look for bugs more. I would love just to have a point release of KDE that simply just fixes as many bugs and compiler warnings as possible in the core. A sort of polishing release, I think this really should be policy to do every so often.
I am carefully considering getting back into KDE dev, I use to port to Solaris and make it build with Sun Studio.
I am considering Clang/LLVM on FreeBSD these days.
Edward.
Post a Comment